run-43/checkpoint-860/trainer_state.json

{
  "best_metric": 0.8166666666666667,
  "best_model_checkpoint": "Bert_tinybert-distilled/run-43/checkpoint-688",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 860,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.9297363758087158,
      "learning_rate": 0.00018220163554157572,
      "loss": 0.5109,
      "step": 172
    },
    {
      "epoch": 1.0,
      "eval_AUC": 0.8338702222222223,
      "eval_BAC": 0.7526666666666666,
      "eval_MCC": 0.5054631978951394,
      "eval_PR_AUC": 0.7906860159456152,
      "eval_accuracy": 0.7526666666666667,
      "eval_loss": 0.45379385352134705,
      "eval_runtime": 6.2406,
      "eval_samples_per_second": 240.361,
      "eval_steps_per_second": 7.531,
      "step": 172
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.4736577272415161,
      "learning_rate": 0.00015617283046420776,
      "loss": 0.4701,
      "step": 344
    },
    {
      "epoch": 2.0,
      "eval_AUC": 0.8569777777777778,
      "eval_BAC": 0.7826666666666666,
      "eval_MCC": 0.5934186315088451,
      "eval_PR_AUC": 0.8041841872474981,
      "eval_accuracy": 0.7826666666666666,
      "eval_loss": 0.4416643977165222,
      "eval_runtime": 6.2534,
      "eval_samples_per_second": 239.869,
      "eval_steps_per_second": 7.516,
      "step": 344
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.348161220550537,
      "learning_rate": 0.0001301440253868398,
      "loss": 0.4558,
      "step": 516
    },
    {
      "epoch": 3.0,
      "eval_AUC": 0.8637493333333333,
      "eval_BAC": 0.804,
      "eval_MCC": 0.6085541718553057,
      "eval_PR_AUC": 0.8226621073554637,
      "eval_accuracy": 0.804,
      "eval_loss": 0.4197617471218109,
      "eval_runtime": 6.2546,
      "eval_samples_per_second": 239.825,
      "eval_steps_per_second": 7.515,
      "step": 516
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.7102890014648438,
      "learning_rate": 0.00010411522030947184,
      "loss": 0.4473,
      "step": 688
    },
    {
      "epoch": 4.0,
      "eval_AUC": 0.8772622222222222,
      "eval_BAC": 0.8166666666666667,
      "eval_MCC": 0.6370595622544093,
      "eval_PR_AUC": 0.8431382446127011,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.40485522150993347,
      "eval_runtime": 6.2302,
      "eval_samples_per_second": 240.762,
      "eval_steps_per_second": 7.544,
      "step": 688
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.8933205604553223,
      "learning_rate": 7.808641523210388e-05,
      "loss": 0.4369,
      "step": 860
    },
    {
      "epoch": 5.0,
      "eval_AUC": 0.8801617777777779,
      "eval_BAC": 0.8153333333333334,
      "eval_MCC": 0.6421162229151887,
      "eval_PR_AUC": 0.8430902218812605,
      "eval_accuracy": 0.8153333333333334,
      "eval_loss": 0.4066833555698395,
      "eval_runtime": 6.2576,
      "eval_samples_per_second": 239.71,
      "eval_steps_per_second": 7.511,
      "step": 860
    }
  ],
  "logging_steps": 500,
  "max_steps": 1376,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2251888650000.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.788436760671977,
    "learning_rate": 0.00020823044061894368,
    "num_train_epochs": 8,
    "temperature": 3
  }
}